Abstract:Background: Detecting event triggers in biomedical texts, which contain domain knowledge and context-dependent terms, is more challenging than in general-domain texts. Most state-of-the-art models rely mainly on external resources such as linguistic tools and knowledge bases to improve system performance. However, they lack effective mechanisms to obtain semantic clues from label specification and sentence context. Given its success in image classification, label representation learning is a promising approach to enhancing biomedical event trigger detection models by leveraging the rich semantics of pre-defined event type labels. Results: In this paper, we propose the Biomedical Label-based Synergistic representation Learning (BioLSL) model, which effectively utilizes event type labels by learning their correlation with trigger words and enriches the representation contextually. The BioLSL model consists of three modules. Firstly, the Domain-specific Joint Encoding module employs a transformer-based, domain-specific pre-trained architecture to jointly encode input sentences and pre-defined event type labels. Secondly, the Label-based Synergistic Representation Learning module learns the semantic relationships between input texts and event type labels, and generates a Label-Trigger Aware Representation (LTAR) and a Label-Context Aware Representation (LCAR) for enhanced semantic representations. Finally, the Trigger Classification module makes structured predictions, where each label is predicted with respect to its neighbours. We conduct experiments on three benchmark BioNLP datasets, namely MLEE, GE09, and GE11, to evaluate our proposed BioLSL model. Results show that BioLSL has achieved state-of-the-art performance, outperforming the baseline models. Conclusions: The proposed BioLSL model demonstrates good performance for biomedical event trigger detection without using any external resources. This suggests that label representation learning and context-aware enhancement are promising directions for improving the task. The key enhancement is that BioLSL effectively learns to construct semantic linkages between the event mentions and type labels, which provide the latent information of label-trigger and label-context relationships in biomedical texts. Moreover, additional experiments on BioLSL show that it performs exceptionally well with limited training data under the data-scarce scenarios.

An improved RL-based framework for multiple biomedical event extraction via self-supervised learning

Efficient multiple biomedical events extraction via reinforcement learning

Joint Event Extraction Based on CNN-BiGRU and Attention Mechanism

Improve Event Extraction via Self-Training with Gradient Guidance

A biomedical event extraction method based on fine-grained and attention mechanism

Biomedical event causal relation extraction with deep knowledge fusion and Roberta-based data augmentation

Effective type label-based synergistic representation learning for biomedical event trigger detection

A Unified Active Learning Framework for Biomedical Relation Extraction.

Bidirectional Long Short-Term Memory with CRF for Detecting Biomedical Event Trigger in FastText Semantic Space

Relation Extraction in Biomedical Texts: A Cross-Sentence Approach

A pre-training and self-training approach for biomedical named entity recognition

Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks

A biomedical events extracted approach based on phrase structure tree

BioREx: Improving Biomedical Relation Extraction by Leveraging Heterogeneous Datasets

A general approach for improving deep learning-based medical relation extraction using a pre-trained model and fine-tuning

Generative Biomedical Event Extraction with Constrained Decoding Strategy

Towards Bootstrapping Biomedical Named Entity Recognition Using Reinforcement Learning

Research on Medical Text Parsing Method Based on BiGRU-BiLSTM Multi-Task Learning

Pipelined Biomedical Event Extraction Rivaling Joint Learning

Using distant supervision to augment manually annotated data for relation extraction

A tree-like structured perceptron for transition-based biomedical event extraction